Issues with LFO controlling LED (help me debug?)

[antiuser]

I've been trying to break down the building blocks of optical tremolo/vibrato-type circuits in order to have a better grasp of the basics, and I'm having a hell of a time with this LFO (which is used on the Phozer and on the Magnavibe, among others).



I can get it to oscillate fine, and the LED will blink, but it's super unstable. If I turn the rate pot up too high, the LED stays on constantly and doesn't return to blinking when I back it down. Not until I power the circuit off and on again. I've tried increasing the resistance between the pot and the capacitors, but that doesn't seem to help.

It seems like it's a grounding issue, because sometimes if I just touch parts of the circuit on my breadboard once the rate pot has been turned down, the LED goes back to blinking. I've looked high and low on this forum and others but haven't been able to find anyone else having the same issue. I've also tried different transistors (3904, MPSA18, BC547, 5088). Pretty much every NPN I have at home. Only the 2N2222 and BC547 seem to make it blink, but both exhibit the same instability.

[antiuser]

BTW that was just a rough schematic, in my haste I forgot to include the 1k resistor I have between the LED and the voltage rail, and got one of the caps backwards.

[PRR]

> If I turn the rate pot up too high, the LED stays on
> I forgot to include the 1k resistor I have between the LED and the voltage rail

The 1K is of course essential.

The circuit needs a raw gain over 27 (29?) to oscillate, and more to start-up reliably.

If it just-starts, and then RATE is turned anywhere near 1K, gain falls, it drops out of oscillation. You "need" a stopper resistor so the user can't go too far. (Not sure why increasing RATE does not re-start it.)

The Ideal Amplifier phase-shift oscillator can be computed. The 1-BJT PSO "can" be computed but you really have to work at it. Even desk-checking a design is not easy.

A 1-BJT amp's gain is limited by power voltage. Sticking the LED in like that "wastes" 1.5+V (more for non-Red LEDs).

My hasty-look suggests the ratio 1K 2.2Meg is non-optimal. However the 2M2 is (supposed to be) the right value for the C-R-C-R-C-R network (after Miller bends it). And the 1K is a reasonable value to light an LED. But is the ratio right for the transistor current gain??

My inclination is to separate the oscillator and the LED driver functions.

[antiuser]

Thanks for that! It seems to be working in a pretty stable way right now. The rate limiting resistor worked out to be 1.5k. I ended up switching the caps from 1uF to 2.2uF because it wasn't going slow enough with the rate pot bottomed out.